Research Interests

Motivated reasoning, selective exposure, misinformation, judgment & decision making, cognitive biases, scientific and economic reasoning, personality & individual differences, religious cognition, evolutionary psychology, public choice, electoral systems, quantitative methods, agent-based models and computer simulations, pedagogy


Individual Differences in Selective Exposure to Attitude-Congruent Political Information: Intuition, Faith, and Social Environment
When seeking political information, people are motivated to selectively seek information that will support their prior beliefs or attitudes rather than information that will challenge them. However, there may be differences in the degree to which individuals engage in such selective exposure. I seek to identify dispositional or environmental variables that may influence the development of such individual differences. I use controlled information-search tasks on a controversial political issue to measure the relative frequency with which subjects choose to read arguments that are congruent with their prior attitudes on that issue. In the first chapter, I show that the preference for reading attitude-congruent information is stronger among individuals who rely more on automatic or intuitive thought processes rather than effortful reflection, as measured by the Cognitive Reflection Test. In the second chapter, I investigate the effects of religious faith. Some religions explicitly teach the importance of maintaining one’s beliefs, and I theorize that habits of selective exposure to information congruent with one’s religious beliefs could, as a side effect, lead to habits of selective exposure in other contexts such as political information seeking. I find that, in an information-search task on a non-religious political issue, the preference for reading attitude-congruent arguments is correlated with scores on a self-report scale of rigid religious conviction and can be increased by priming people to think of religion. The third chapter investigates the effects of having a politically homogeneous or heterogeneous social environment. To enable stronger causal inference, I study residents of student housing, a situation in which many residents live with people they have not chosen on the basis of political similarity. I do not find clear and consistent evidence that the preference for reading attitude-congruent information is stronger or weaker among residents who live with a politically similar roommate, though I do find such an effect among a sub-group: Republican-leaning residents. I also do not find strong evidence that priming people to think of the members of their social networks as more politically similar or dissimilar to their selves affects the tendency to read congruent information.

Full text: Cragun dissertation.pdf


Rigid Religious Faith Promotes Selective Exposure to Attitude-Congruent Political Information
Political Behavior, 2020
When seeking political information, people are motivated to selectively seek information that is congruent with their prior attitudes. However, some individuals may do so more than others, and not much is known about what factors affect such individual differences. Rigid religious faith is one variable that may promote selective exposure. Messages of the importance of rigid faith --- the idea that religious beliefs must be held firmly and not doubted --- could encourage a habit of selective exposure to information that supports existing religious beliefs. As a side effect, this habit of selective exposure might be applied outside the context of religion. In this study, an information-search task on a non-religious political issue is used to demonstrate that subjects prefer to read a greater number of arguments that are congruent with their prior attitudes on the issue, and this effect of prior attitudes on information-search behavior is found to be stronger among individuals who have rigid religious convictions. A scrambled-sentence task is used to prime half the subjects with religious concepts prior to completing the information-search task. This experiment demonstrates that increased salience of religious faith causes an increase in selective exposure to attitude-congruent political information.

More details:

Observational results: Religious faith and selective exposure

Experimental results: Religious faith and selective exposure

Other current projects

Evolution of One-Shot Cooperation Under Uncertainty: A Revised Agent-Based Model
Cooperation for mutual benefit is a fundamental aspect of politics, and behavioral economics research has provided much knowledge about the degree to which humans tend to cooperate and the conditions under which they do so. Although it is often not optimal, from a traditional economic (rational egoist) perspective, to cooperate when in a one-shot (non-repeated) interaction, experimental research has shown that many individuals do cooperate even in one-shot interactions. This may be because the typical experimental situation in which subjects are told with certainty that they are in a one-shot interaction is unlike the situations humans often encounter in real life. In real interactions, an individual would have a belief about whether the current interaction is a one-shot or repeated interaction, but those beliefs could sometimes be wrong. The opportunity cost of failing to cooperate when an interaction turns out to be a repeated interaction may be greater than the cost of cooperating when an interaction turns out to be a one-shot interaction, so a tendency to cooperate even when the individual believes the interaction is more likely to be a one-shot interaction could be beneficial in the long run. I use a computer simulation to show the evolution of this behavioral trait. Over thousands of generations of simulated agents interacting with each other, producing offspring in proportion to their level of success in those interactions, and passing their behavioral tendencies to their offspring, the tendency to cooperate even when they believe a one-shot interaction is more likely becomes common in the population. This model is based on a model previously published by Delton et al (2011) but includes some important modifications. More details, and an interactive web-based interface for the simulation, can be found here:

The importance of interactive control variables when testing interactive hypotheses
In political science, we often want to know how the effect of X on Y depends on the value of M, a moderating variable. This is sometimes measured using a regression model of Y that includes a multiplicative interaction between M and X. In many studies, the primary hypothesis of interest is that M affects the effect of X on Y. In such situations the outcome of interest is not just the value of Y but the effect of X on Y. Sometimes researchers include extra ``control" variables in the model because they are concerned that one of these other variables may be a confounding variable. However, many researchers do not think carefully enough about how a potential confound might occur when testing an interactive hypothesis. They fail to realize that adding a control variable C, without also adding a multiplicative interaction between C and X, will only control for simple additive effects of C on Y and will not control for any confounding effects C may have on the outcome of primary interest (the outcome of primary interest being the effect of X on Y). A more detailed explanation of the problem can be found here: I am currently working on simulations to determine in what situations the failure to include an interactive control variable is most problematic, and I am conducting a review of the political science literature to determine how often political scientists attempt to control for potential confounding variables when testing an interactive hypothesis but fail to include interactive control variables.

Does anticipation of a coming vaccine affect willingness to comply with prevention measures now?
Does telling people that a COVID vaccine will soon be available make people more likely or less likely to engage in risky behaviors now? I conducted a study to answer this question. My results can be seen here: